CN1187251C - Device for managing and controlling operation of elevator - Google Patents

Abstract

The elevator operation management and control system of the present invention includes a traffic data collection part for collecting traffic data for obtaining elevator passenger traffic volume; a traffic volume calculation part for calculating the traffic volume according to the traffic data collected by the traffic data collection part A traffic volume calculation section of volume; a traffic flow calculation section for calculating a traffic flow estimated value of the elevator passengers moving between floors based on the traffic volume calculated by the traffic volume calculation section; control parameter setting a section for setting a control parameter for controlling the operation of the elevator based on the traffic flow estimated value calculated by the traffic flow calculation section; and an operation control section for setting a control parameter based on the control parameter setting section by the The control parameters set are used to control the operation of the elevator, so that it does not need to prepare and store a large number of traffic flow patterns and combinations of traffic volumes obtained according to the traffic flow patterns in advance, and can immediately learn from the hitherto observed The traffic volume data calculates the traffic flow estimated value, and by setting the control parameters of the group management control corresponding to the traffic flow estimated value, the elevator group management control is performed.

Description

Elevator operation management and control system

technical field

The present invention relates to elevator operation management and control systems.

Background technique

FIG. 7 is an explanatory diagram showing the basic principle for estimating the traffic flow of a prior art traffic control system such as that described in JP-A-7-309546 and whose control purpose is particularly for vehicles including multiple elevators .

In Fig. 7, reference numeral 11 represents traffic volume data, and it includes quantitative information such as the number of people who take the elevator on each floor and the number of people who leave the elevator on each floor, and reference numeral 13 represents instructions such as by volume, time zone (timezone), direction, etc. The elements represent the generation and movement of elevator passengers and the number 12 represents the multi-level for traffic flow 13 from input traffic data 11 derived from preset relationships between traffic volumes and traffic flow patterns. Neural network (control neural network).

Assuming that in a certain building, the number of elevator passengers who take the elevator on the i-th floor and leave the elevator on the j-th floor within a predetermined period of time, that is, the number of elevator passengers moving from the i-th floor to the j-th floor is Tij , the traffic flow in that time period in that time period can be represented as follows:

Traffic flow: T = (T12, T13, ..., Tij, ...) ... (1)

Then, the traffic volume data generated and observed through this traffic flow can be expressed as follows:

Traffic volume data: G = (p, q)                                                                                                                                                                                            ... (2)

Among them, p is the number of people taking the elevator on each floor, and q is the number of people leaving the elevator on each floor.

Thus, the traffic flow is the flow of this traffic, and the traffic volume is an easily observable quantity that can be found from the traffic flow.

In addition, when the observable control result is set to E, in addition to the traffic volume data, the control result E can also be expressed as follows:

Control results: e = (r, y, m) ... (3)

where r is the distribution of response time to hall calls, y is the distribution of prediction misses per floor and m is the distribution of times when the elevator (car) is full and passes each floor.

Since it is difficult to find the traffic flow T precisely and directly from the traffic data G which does not contain information on the direction of movement of the elevator passengers in the target time period, an approximate method is used here to find the traffic flow.

First, a large number of hypothetical traffic flow patterns in the building are prepared in advance, and then fixed control parameters are found through simulation, and the traffic volume data G and control results E generated when each traffic flow pattern is controlled. Thus, several relationships between "traffic volume and traffic flow pattern" and "traffic flow pattern and control result" can be obtained.

Next, a neural network is used to represent the relationship between "traffic volume and traffic flow pattern". Then, for example, a multilayer neural network 12 as shown in FIG. 7 is prepared, and the traffic data 11 is provided to the input side, and the traffic flow pattern 13 that generates the traffic data 11 is provided to the output side, as a so-called teacher for the neural network to learn. (teacher) data.

As a result, when a certain traffic volume data is input, the neural network 12 outputs a traffic flow pattern which is the closest to the one which can generate the input traffic volume data among the various traffic flow patterns prepared in advance.

Therefore, by preparing and having the neural network 12 learn a sufficient number of traffic flow patterns in advance, the neural network 12 selects and outputs the generated arbitrary traffic flow pattern with respect to the traffic flow data from the relationship of "traffic volume and traffic flow pattern" learned so far. volume of traffic flow, or at least traffic flow that is very close to it.

When the same traffic flow data is generated according to a plurality of different traffic flow patterns, since under fixed control parameters, when the traffic flow is different, the control results are different from each other, so the neural network 12 can select the traffic flow pattern, which allows relationship between flow patterns and control outcomes" to obtain specific control outcomes from traffic flow patterns that generate the same traffic volume data.

In addition, since the control parameters can be set so that the pre-prepared traffic flow pattern can obtain the best control results by means of pre-simulation, etc., when the traffic flow can be estimated according to the traffic volume data, the neural network 12 can set the optimal control parameters .

In this prior art, determining the accuracy of the traffic flow estimation depends on how many combinations between the traffic flow pattern and the traffic volume obtained from the traffic flow pattern can be prepared in advance. However, there is a problem that since preparing and storing in advance all kinds of traffic flow patterns and combinations of traffic volumes obtained from the traffic flow patterns requires a huge storage capacity, and it cannot efficiently allocate appropriate elevators corresponding to the current service state, so This is not practical.

The technique described in JP-B-62-36954 also has a problem, that is, since although it can analyze what kind of traffic flow occurred in the past, it cannot estimate in real time what kind of traffic flow occurred at that time while controlling elevator operation management, so It cannot effectively allocate the appropriate elevator corresponding to the current service state.

Therefore, it is an object of the present invention to solve this problem by providing an elevator operation management and control system capable of estimating traffic flow from observed traffic data in real time and capable of performing elevator operation management and control corresponding to the estimated traffic flow .

Contents of the invention

The present invention provides an elevator operation management and control system, comprising:

The traffic data collection part is used to collect traffic data to obtain the traffic volume of elevator passengers;

a control parameter setting section for setting control parameters for controlling the operation of the elevator; and

an operation control section for controlling the operation of the elevator according to the control parameter set by the control parameter setting section,

It is characterized in that the elevator operation management and control system also includes:

a traffic volume calculation section for calculating the traffic volume based on the traffic data collected by the traffic data collection section;

a traffic flow calculation section for calculating in real time an estimated value of traffic flow of the elevator passengers moving between floors based on the traffic volume calculated by the traffic volume calculation section;

a traffic data analysis section for analyzing movement of passengers according to said traffic data;

an estimating function constituting section for constituting an estimating function of said traffic flow calculating section based on a result analyzed by said traffic data analyzing section;

Wherein, the control parameter setting section sets a control parameter for controlling the operation of the elevator based on the traffic flow estimated value calculated by the traffic flow calculation section.

In the elevator operation management and control system as described above, the traffic flow calculation section is formed by a neural network, wherein the traffic volume of the elevator passengers is set on the input side of the neural network, and the traffic volume of the elevator passengers is set on the output side. The traffic flow of the elevator passengers.

The elevator operation management and control system as described above, further comprising a teacher data generation section, wherein the teacher data generation section is used to generate the neural network for learning based on the traffic data collected by the traffic data collection section The estimation function constituting section calculates the traffic flow estimation value of the traffic flow calculating section by providing the neural network with learning based on the teacher data generated by the teacher data generating section.

In the elevator operation management and control system as described above, the estimating function constituting part sets a value as an index of the estimating accuracy based on the error of two squares of each corresponding element of the traffic flow data, wherein the two squares The employed teacher data and the traffic flow estimated value calculated by the traffic flow calculating section based on the traffic volume data of the teacher data correspond respectively.

In the elevator operation management and control system as described above, the teacher data generating section generates the teacher data based on the traffic data collected by the traffic data collecting section within a predetermined period of time.

In the elevator operation management and control system as described above, the traffic flow calculation section calculates the traffic flow estimation value as a ratio of the traffic volume of elevator passengers moving between the target floors to the entire traffic volume.

In the elevator operation management and control system as described above, the operation control section performs operation control by group management control.

Description of drawings

Fig. 1 is an explanatory diagram of the elevator operation management and control system of the present invention.

Fig. 2 is an explanatory diagram of the elevator operation management and control system of the present invention.

Fig. 3 is an explanatory diagram of the elevator operation management and control system of the present invention.

Fig. 4 is an explanatory diagram of the elevator operation management and control system of the present invention.

Fig. 5 is an explanatory diagram of the elevator operation management and control system of the present invention.

Fig. 6 is an explanatory diagram of the elevator operation management and control system of the present invention.

Fig. 7 is an explanatory diagram of a prior art traffic device control system.

Detailed ways

first embodiment

Next, a first embodiment of the present invention will be described using the drawings.

Fig. 1 is an explanatory diagram of the basic principle of traffic flow estimation of the elevator operation management and control system of the present invention. Here, the principle is explained by exemplifying a case where a plurality of elevators are operated using group supervisory control.

In Fig. 1, the traffic volume data 11 includes quantitative information, such as, each direction (UP/DOWN) of each floor, the number of people who take the elevator and the number of people who leave the elevator, and express the traffic with OD (originating/destination) data Stream 13, wherein the data represent the traffic of elevator passengers moving from one floor to another target floor in the total traffic. A multilayer neural network (controlling neural network) 12 estimates traffic flow data 13 from input traffic volume data 11 .

When it is assumed here that in a certain building, within a predetermined period of time, the number of elevator passengers who take the elevator from the i-th floor and leave the elevator on the j-th floor, that is, the number of elevators moving from the i-th floor to the j-th floor The OD data of the number of passengers, when set as TFij, can represent the traffic flow in the building as follows in the same way as the prior art example described above, because it is a collection of those OD data:

Traffic flow: TF = (TF ₁₂ , TF ₁₃ , ..., TFij, ...) ... (4)

Furthermore, the observable traffic volume data generated by such a traffic flow is represented as follows:

Traffic volume data: G=(ON up(f1), ON dn(F1), OFF up(f1), OFF dn(f`1))

ON up(f1): On the f1 floor, the number of people taking the elevator in the upward direction,

ON dn(F1): On the f1 floor, the number of people taking the elevator in the downward direction,

OFF up(f1): On the f1 floor, the number of people leaving the elevator in the upward direction,

OFF dn(F1): On the f1 floor, the number of people leaving the elevator in the downward direction ...(5)

Generally, although the traffic volume T shown in expression (5) can be found from the traffic flow data G, it is difficult to find the precise traffic flow G from the traffic volume data T instead, where the traffic flow data G contains the elevator passenger The direction of movement and the information of the target time period shown in expression (4).

Thus, according to the present invention, in addition to the daily group management and control, the neural network is also used to find the floor as each floor based on the past table of each traffic flow data about how many elevator passengers moved from which floor to which floor in the target time period. The traffic volume of the number of elevator passengers between elevators, and the map of the traffic volume defined by the traffic flow data is represented by the neural network. Then, by using the learning result of this neural network in the control group management process, the traffic flow G is approximately found from the traffic volume data by using the inverse mapping of this map.

Thus, for example, the neural network learns the relationship between the traffic flow and the traffic volume calculated from it, eg after the end of the daily control. When the traffic volume data is given from the input side, the neural network is made to learn, and in this case the traffic flow is extracted from the output side, when a certain traffic volume data is input as the general quality of the neural network, the neural network can output the same traffic Traffic flow corresponding to volume data. That is, the neural network can acquire the ability to perform mapping that is the inverse of a mapping that defines traffic volumes from traffic flow data.

Although the operation control system performs group management by setting control parameters corresponding to the traffic flow when the traffic flow can be specified, there are multiple control parameters in elevator group management control, such as the number of elevators assigned to congested floors, setting No floors served, forecast of when each elevator will reach a particular floor, weighting of each evaluation index during call assignment, etc.

However, when a traffic flow can be specified, a method such as simulation can be used to evaluate the control result under limited control parameters, and the optimum value of the control parameter for each traffic flow can be set. That is, when the traffic flow can be estimated, the optimal value of the control parameter can be automatically set.

Next, as an embodiment of the present invention, an elevator operation management and control system for controlling a plurality of elevator groups based on the traffic flow estimated by the above basic principle can be explained using FIG. 2 .

Fig. 2 is a block diagram showing the structure of a group management and control system as an example of the elevator operation management and control system of the present invention. In FIG. 2, reference numerals (31 to 3n) denote hall call buttons provided at the halls of each floor. When an elevator passenger manipulates at least any one of the hall call buttons 31 to 3n, a hall call is output from the manipulated hall call button to the group management control unit 1, whereby the group management control unit 1 performs group management control.

Each of the elevator controllers 21 to 2m operates each elevator, such as running, stopping, and door opening/closing, according to control commands of the group supervisory control unit 1 .

Here, the group management control unit 1 includes a traffic data collection section 1A to collect traffic data such as actions of each elevator and generated calls, a traffic volume calculation section 1B to calculate traffic volume based on the collected traffic data, An estimating section 1C as a traffic flow calculating section to calculate a traffic flow estimated value in real time based on calculated traffic data, a teacher data generating section to generate teacher data for learning a neural network by analyzing movement of elevator passengers based on traffic data 1D. An estimating function constituting section 1E for constituting the function of the traffic flow estimating section 1C for calculating an estimated value of traffic flow by learning a neural network based on the teacher data generated by the traffic data generating section 1D. The estimated value of the traffic flow estimated by the estimating section 1C, the control parameter setting section 1F for controlling the control parameters of the elevator group and the operation control section 1G for controlling the group management according to the preset control parameters are set.

Here, the above-mentioned traffic data includes not only data used to calculate the traffic volume but also data used to analyze the movement of elevator passengers to estimate traffic flow, such as signals such as calls made by elevator passengers, such as stop, up, Elevator operation information such as down, the number of people boarding/leaving the elevator, information about elevators such as changes in load and target time period.

Referring to Fig. 3, the specific operation of the elevator group management control will be explained especially as the operation of this embodiment.

Fig. 3 is a flowchart schematically showing group management control.

First, the traffic data collection section 1A collects traffic data such as elevator behavior such as stop and run, number of people boarding/leaving the elevator, elevator calls, hall calls, and called elevators in real time (step ST10).

Next, the traffic amount calculation section 1B calculates traffic amount data G based on the traffic data collected by the traffic data collection section 1A (step ST20). The calculation of the traffic volume can be realized by causing the traffic volume calculation section 1B to count the number of people who got on/off the elevator in the past 5 minutes, for example, every 1 minute.

Next, the traffic flow estimation section 1C calculates a traffic flow estimation value in real time based on the traffic volume data calculated by the traffic volume calculation section 1B (step ST30). Here, referring to FIG. 4, the traffic flow estimation operation in step S30 can be explained.

The calculated traffic volume data G is input to the neural network 12 shown in FIG. 1 (step ST31). At this time, the values of the respective element data ON up(f1), ON dn(F1), OFF up(f1) and OFF dn(f`1) of the traffic volume data G as shown in the expression (2) Input to each neuron in the input layer of the neural network 12 . Therefore, the number of neurons in the input layer is 4×Z (Z is the number of layers in the building).

Here, the neural network 12 performs known network calculation (step ST32) and outputs the estimated value of traffic volume found through the calculation in real time.

In this case, the output value of each neuron in the output layer of the neural network 12 is set as the estimated value of each element of the traffic flow data TF in Expression (4). That is, by setting the output value of each neuron of the output layer as TF11, the output value of the second neuron as TF12, . . . , an estimated value of traffic flow data can be obtained as OD data. Therefore, the number of neurons in the output layer is Z ² .

Note that the number of neurons in the middle layer can be arbitrarily set corresponding to each case.

Furthermore, by dividing the building into several zones, each zone can describe the traffic flow and volume. In this case, Z above is the number of zones.

Now, returning to the explanation of FIG. 3, when the traffic flow estimation data is obtained in real time by the neural network 12 in step ST30, then the control parameter setting section 1F sets control parameters corresponding to the traffic flow estimated by the neural network 12. (step ST40).

Next, the operation control section 1G executes elevator group management control based on the control parameters set by the control parameter setting section 1F (step ST50).

By the way, the function to estimate the traffic flow based on the traffic volume data realized by the neural network 12 during the daily group management control can be constituted by repeatedly correcting the estimation function described below.

That is, for example, the correction of the traffic flow estimation function realized by the neural network 12 is intermittently performed separately from the daily group management control (step ST60). Correction of the estimated function may be performed after daily control is completed, or at predetermined time intervals such as every week.

Correction of the estimation function can be achieved by causing the neural network 12 to learn the relationship between traffic flow and traffic volume, and by causing the neural network 12 to improve the capability of a particular traffic flow estimation function beyond the capability of the traffic flow estimation function obtained last time, wherein according to The above-described traffic flow and traffic volume are calculated from traffic flow data and traffic volume data found from traffic data obtained between the last execution of the correction estimation function and the execution of the correction estimation function this time.

The procedure to correct the estimation function (step ST60) is described with reference to FIG. 5 .

FIG. 5 is a flowchart showing a function to correct traffic flow estimation.

Data stored for correcting the estimation function is fetched from the traffic data under group management control collected in step ST10 (step ST61).

Regarding the traffic data used to correct the estimation function, it is not necessary to store all collected data as data used for correction. Predetermined data of about 5 minutes can be set as a unit, and a predetermined amount of data, such as several data per time period, such as office hours and normal hours in which characteristic traffic occurs, can be stored for use in correcting the estimation function.

Next, the teacher data generating section 1D analyzes the traffic data for correcting the estimation function to generate so-called teacher data for learning the neural network 12 (step ST62).

Here, the teacher data includes a combination of traffic volume data and traffic flow data respectively from the traffic data. Here, the traffic volume data is found in the form of expression (5) in the same way as the procedure of the above-mentioned step ST20 according to the number of people boarding/leaving each elevator. Traffic flow data can be found in the form of expression (4). Referring to Figure 6, the process of finding them is further explained.

The sequence of operations of an elevator from when it starts running either up or down to when it reverses its course is called a scan. For example, assume that, during the target time period, the number of people who stop at a floor and take/leave an elevator is 1F (3 people get on) → 3F (2 people leave) → 4F (1 person gets on) in the upward scan → 6F (1 person leaves) → 10F (10 people leave), as shown in Figure 6.

In this case, the two people who got off the elevator at 3F can be specified as the people who got on the elevator from IF. However, the boarding floors of the elevator passengers leaving the elevator at 6F and 10F cannot be specified.

As a result, the unspecified number of people who have left the elevator are assigned equally to the movement combination of the elevator passengers. That is, in this case, two unspecified persons 1F→6F (0.5 persons), 4F→6F (0.5 persons), 1F→10F (0.5 persons) and 4F→10F (0.5 persons) are assigned as follows.

Next, each region transforms this data. When, in the example of FIG. 6 , IF is set as the first area, 2F to 6F is set as the second area, and 7F to 10F is set as the third area, the traffic flow data is expressed as OD as in the following expression (6). (origin/destination) data:

TF12＝2.5 (1F→3F (2 persons) and 1F→6F (0.5 persons))

TF13＝0.5(1F→10F(0.5 people))

TF22＝0.54F→6F (0.5 people)

TF23＝0.5(4F→10F(0.5 people)) ...(6)

By calculating and integrating the above process per elevator and per scan, traffic flow data can be found, which reflects the movement information about the passengers of each elevator in the target time period.

Then, the neural network 12 is made to learn to adjust the neural network 12 by using the traffic volume data and data flow data obtained for each stored traffic data as teacher data (step ST63).

A known so-called backpropagation method is used for learning the neural network 12 .

Next, the estimation accuracy of the traffic flow is checked. As an index of the estimation accuracy, the sum of the squared errors of the corresponding elements of the employed teacher data and the traffic flow estimated value calculated by the neural network 12 from the traffic volume data of the teacher data is used (step ST64).

That is, the errors E respectively found by the following expression (7) with respect to all teacher data are summed up and the total value is set as an index of the estimation accuracy. It can be considered that the smaller the total value, the better the estimation accuracy.

E＝∑(TFij- TF ij) ²

TFij = each element value of the traffic flow data of the teacher data

TF ij = each element value of the traffic estimate calculated from the traffic data of the teacher data

...(7)

Next, the estimation function constituting section 1E compares the total value of errors E found by using expression (7) with the total value E of errors found by using expression (7) in the process of correcting the estimation function performed last time Comparison (step ST65).

Then, when the estimation accuracy improves (YES in step ST65), the estimation function constituent part IE registers the neural network adjusted in step S63 (step ST67), otherwise, when the accuracy fails to improve (in step S765 No) then return to the previous neural network to log in.

The neural network 12 and the traffic flow estimating section 1C are always kept in good condition, and the accuracy of the estimated traffic flow is kept well by performing correction of the traffic flow estimation function in addition to the normal group management control.

Therefore, the above-described embodiment does not need to prepare and store a large number of combinations of traffic flow patterns and traffic volume obtained according to the traffic flow patterns in advance, immediately calculate traffic flow estimation values from traffic volume data observed so far, and calculate traffic flow by setting The control parameters of the group management control corresponding to the flow estimation value can be used for elevator group management control.

Furthermore, since the input data does not contain any estimates and is immediately observable traffic volumes, traffic flow can be calculated with high accuracy and more precisely estimated. In addition, since the present embodiment is so arranged that the relationship between traffic volume and traffic flow is generated by the neural network, and the estimation function is constituted and corrected by making the neural network learn the analysis results of the traffic data, it does not need to store a large amount of data in advance. Data to connect the relationship between the two with a large amount of logic, and can be reduced to the program and storage area required to connect the two for calculation.

In addition, since the traffic flow estimated value estimated by the traffic flow estimating section can be estimated by Accuracy is maintained so well that this embodiment allows the elevator operation management and control system to follow the movement of elevator passengers per building hit, where the changes are building and time period dependent.

In addition, learning is performed by using non-stationary traffic flow data to calculate an index of estimation accuracy of the traffic flow estimation portion as teacher data, so there is no need to worry about deterioration of estimation accuracy by the estimation function component.

The present invention allows the neural network to be tuned because it estimates the traffic flow corresponding to the time period at each predetermined time period by using the teacher data, thus it allows a more accurate estimation of the traffic flow corresponding to the time period, rather than using whatever Time periods are consistently estimated for the calculated portion of the traffic flow.

In addition, in the entire traffic volume, the traffic flow calculation section calculates the traffic flow estimation value as the traffic volume of the elevator passengers moving between the target floors, so the movement of the elevator passengers in the building is expressed without error.

In addition, the present invention is not only effective in controlling the operation management of one elevator, but also allows complicated elevator operation management to be used for so-called group management control for performing optimum operation control by assigning calls to a plurality of elevators mutually.

As described above, the elevator operation management and control system of the present invention can be suitably used.

Claims (7)

1. A lift operation management and control system, comprising: The traffic data collection part is used to collect traffic data to obtain the traffic volume of elevator passengers; a control parameter setting section for setting control parameters for controlling the operation of the elevator; and an operation control section for controlling the operation of the elevator according to the control parameter set by the control parameter setting section, It is characterized in that the elevator operation management and control system also includes: a traffic volume calculation section for calculating the traffic volume based on the traffic data collected by the traffic data collection section; a traffic flow calculation section for calculating in real time an estimated value of traffic flow of the elevator passengers moving between floors based on the traffic volume calculated by the traffic volume calculation section; a traffic data analysis section for analyzing movement of passengers according to said traffic data; an estimating function constituting section for constituting an estimating function of said traffic flow calculating section based on a result analyzed by said traffic data analyzing section; Wherein, the control parameter setting section sets a control parameter for controlling the operation of the elevator based on the traffic flow estimated value calculated by the traffic flow calculation section. 2. The elevator operation management and control system according to claim 1, wherein said traffic flow calculation part is formed by a neural network, wherein said traffic flow of said elevator passengers is set at an input side of said neural network amount, while the traffic flow of the elevator passengers is set on the output side. 3. The elevator operation management and control system according to claim 2, further comprising a teacher data generation part, wherein the teacher data generation part is used to to generate the teacher data used by the neural network for learning, the estimation function constituting part calculates the The estimated value of the traffic flow. 4. The elevator operation management and control system according to claim 3, wherein the estimation function component sets a value as the estimation accuracy according to the error of two squares of each corresponding element of the traffic flow data , wherein the two squares respectively correspond to the adopted teacher data and the traffic flow estimation value calculated by the traffic flow calculation part based on the traffic volume data of the teacher data. 5. The elevator operation management and control system according to claim 3 or 4, wherein said teacher data generation part generates said teacher data according to traffic data collected by said traffic data collection part within a predetermined period of time. data. 6. The elevator operation management and control system according to claim 5, wherein the traffic flow calculation section calculates the traffic flow estimate by the ratio of the traffic volume of the elevator passengers moving between the target floors to the entire traffic volume value. 7. The elevator operation management and control system according to claim 6, wherein said operation control section performs operation control by group management control.

Images